Tuning the Electronic Structure of CoO Nanowire Arrays by N-Doping for Efficient Hydrogen Evolution in Alkaline Solutions

: Electrochemical hydrogen evolution reactions (HER) have drawn tremendous interest for the scalable and sustainable conversion of renewable electricity to clear hydrogen fuel. However, the sluggish kinetics of the water dissociation step severely restricts the high production of hydrogen in alkaline media. Tuning the electronic structure by doping is an effective method to boost water dissociation in alkaline solutions. In this study, N-doped CoO nanowire arrays (N-CoO) were designed and prepared using a simple method. X-ray diffraction (XRD), element mappings and X-ray photoelectron spectroscopy (XPS) demonstrated that N was successfully incorporated into the lattice of CoO. The XPS of Co 2p and O 1s suggested that the electronic structure of CoO was obviously modulated after the incorporation of N, which improved the adsorption and activation of water molecules. The energy barriers obtained from the Arrhenius relationship of the current density at different temperatures indicated that the N-CoO nanowire arrays accelerated the water dissociation in the HER process. As a result, the N-CoO nanowire arrays showed an excellent performance of HER in alkaline condition. At a current density of 10 mA cm − 1 , the N-CoO nanowire arrays needed only a 123 mV potential, which was much lower than that of CoO (285 mV). This simple design strategy provides some new inspiration to promote water dissociation for HER in alkaline solutions at the atomic level.


Introduction
Due to severe environmental pollution as a result of the excessive consumption of traditional fossil fuels [1][2][3][4], it is desirable to explore clean and sustainable energy sources [5,6]. Hydrogen is of significant importance in sustainable energy systems and in the chemical industry; its advantages include its high energy density as well as being pollution-free [7][8][9]. Electrochemical hydrogen production with electric energy from renewable energy sources, such as wind energy and solar energy, has attracted some attention and has been applied in various hydrogen production technologies. However, the industrial application of water electrolysis is hampered by the performance and cost of HER catalysts. In water electrolysis, HER catalysts can be performed in basic, neutral or acidic solutions [10,11]. Currently, the best HER and oxygen evolution reaction catalysts are still Pt-and Ir/Ru-based materials [12,13], respectively. However, since most of the cost-effective oxygen evolution reaction catalysts on the counter electrode can only work under alkaline condition [14][15][16], developing highly efficient, robust and earth-abundant alkaline HER electrocatalysts is much more desirable [17][18][19].
Over the past few years, a lot of low-cost earth-abundant catalysts have been developed for HER, such as transition metal-based alloys [20][21][22][23][24], nitrides [25,26], carbides [27][28][29], phosphides [30][31][32], sulfides [33][34][35] and polymer catalysts [36][37][38][39][40]. Among them, transition metal nitrides with excellent activity have attracted great attention. However, when compared in acid condition, their HER activities in alkaline media need to be further improved. Due to the sluggish kinetics of water dissociation in alkaline media, the reaction rates of transition metal nitrides for HER in alkaline solutions are about two to three orders of magnitude lower than in acidic solution [41,42]. Previous approaches to solving water dissociation in alkaline solutions focused on interface engineering. Some recent studies based on interface engineering have improved HER performance in alkaline solutions with the enhancement of water dissociation [31,43,44]. However, interface engineering relies on a well-designed interface, with features such as corners, edges and grain boundaries. In addition, the rare metal of Pt is often used to offer an optimal H adsorption energy in these approaches [45,46]. As such, developing new catalysts from a moderate method using earth-abundant elements with efficient water dissociation at atomic scale is highly desirable but challenging.
Tuning the electronic structure by heteroatom doping has been demonstrated to be an effective method to boost water dissociation in alkaline solutions [47][48][49]. CoO is a promising electrocatalyst for HER due to its abundance. However, its high water dissociation energy makes it unsuitable for HER in alkaline solutions [50]. Inspired by the reasons above, a simple and efficient strategy to solve the sluggish water dissociation kinetics of CoO electrocatalysts by N-doping was developed. X-ray diffraction (XRD), element mappings and X-ray photoelectron spectroscopy (XPS) demonstrated that the incorporation of N did not change the crystal structure of CoO. The XPS analysis of the Co 2p and O 1s of CoO and N-CoO proved that the adsorption and activation of water molecules was enhanced significantly after the incorporation of N, which was consistent with the energy barriers obtained from the Arrhenius relationship of the current density at different temperatures. As a result, the N-CoO nanowire arrays exhibited a very high activity, using only 123 mV to achieve the current density of 10 mA cm −2 and the low Tafel slope of 97 mV dec −1 in 1 M KOH. Figure 1 is the schematic illustration for the preparation of the CoO and N-CoO nanowire arrays. Briefly, the CoO and N-CoO nanowire arrays were obtained by annealing the precursor of the Co x (OH) y nanowire arrays under Ar and NH 3 atmospheres, respectively (for details, see Experiment). The Co x (OH) y precursor was prepared by a modified hydrothermal method (for details, see Experiment). Over the past few years, a lot of low-cost earth-abundant catalysts have been developed for HER, such as transition metal-based alloys [20][21][22][23][24], nitrides [25,26], carbides [27][28][29], phosphides [30][31][32], sulfides [33][34][35] and polymer catalysts [36][37][38][39][40]. Among them, transition metal nitrides with excellent activity have attracted great attention. However, when compared in acid condition, their HER activities in alkaline media need to be further improved. Due to the sluggish kinetics of water dissociation in alkaline media, the reaction rates of transition metal nitrides for HER in alkaline solutions are about two to three orders of magnitude lower than in acidic solution [41,42]. Previous approaches to solving water dissociation in alkaline solutions focused on interface engineering. Some recent studies based on interface engineering have improved HER performance in alkaline solutions with the enhancement of water dissociation [31,43,44]. However, interface engineering relies on a well-designed interface, with features such as corners, edges and grain boundaries. In addition, the rare metal of Pt is often used to offer an optimal H adsorption energy in these approaches [45,46]. As such, developing new catalysts from a moderate method using earth-abundant elements with efficient water dissociation at atomic scale is highly desirable but challenging.

Results
Tuning the electronic structure by heteroatom doping has been demonstrated to be an effective method to boost water dissociation in alkaline solutions [47][48][49]. CoO is a promising electrocatalyst for HER due to its abundance. However, its high water dissociation energy makes it unsuitable for HER in alkaline solutions [50]. Inspired by the reasons above, a simple and efficient strategy to solve the sluggish water dissociation kinetics of CoO electrocatalysts by N-doping was developed. X-ray diffraction (XRD), element mappings and X-ray photoelectron spectroscopy (XPS) demonstrated that the incorporation of N did not change the crystal structure of CoO. The XPS analysis of the Co 2p and O 1s of CoO and N-CoO proved that the adsorption and activation of water molecules was enhanced significantly after the incorporation of N, which was consistent with the energy barriers obtained from the Arrhenius relationship of the current density at different temperatures. As a result, the N-CoO nanowire arrays exhibited a very high activity, using only 123 mV to achieve the current density of 10 mA cm −2 and the low Tafel slope of 97 mV dec −1 in 1 M KOH. Figure 1 is the schematic illustration for the preparation of the CoO and N-CoO nanowire arrays. Briefly, the CoO and N-CoO nanowire arrays were obtained by annealing the precursor of the Cox(OH)y nanowire arrays under Ar and NH3 atmospheres, respectively (for details, see Experiment). The Cox(OH)y precursor was prepared by a modified hydrothermal method (for details, see Experiment).   The N-CoO nanowire arrays were prepared by the topotactical transformation of the pre-synthesized Co x (OH) y nanowire arrays using a nitrogen strategy. The XRD patterns of the CoO and N-CoO nanowire arrays in Figure 2a corresponded to the crystal structure of CoO (JCPDS: 01-078-0431), suggesting the precursor of the Co x (OH) y nanowire arrays could convert into CoO and N-CoO nanowire arrays after annealing under Ar and NH 3 atmospheres, respectively. The SEM images of Figures 2b and S1 show the morphology of Co x (OH) y with dense nanowire arrays on the carbon cloth (CC). The SEM images of CoO and N-CoO in Figure 2c,d display the overall sheet-like morphology of the Co x (OH) y nanowire arrays. However, compared to the smooth surface of the CoO nanowire arrays, the surface of the N-CoO nanowire arrays became rough after nitrogen treatment. The TEM image of N-CoO in Figure S2   XPS measurements were carried out to study the electronic structure of the surface elements of CoO and N-CoO nanowire arrays. Figure 3a shows the high-resolution XPS spectra of C 1s, which were used as the reference standard (peak at 284.6 eV) for elements of Co, O and N. As seen in Figure 2b, the element peaks of C 1s, O 1s could be clearly seen in the survey scan spectra of the CoO and N-CoO nanowire arrays. However, compared with CoO, there was a weak peak at around 400 eV, suggesting the existence of the N element in N-CoO. Figure 2c depicts the high-resolution XPS spectra of N 1s, further demonstrating the presence of N in N-CoO, which was consistent with the results of the element mappings. Figure 2d shows the high-resolution spectra of Co 2p; the broad peaks around 780.7 and 796.8 eV were regarded as Co 2p 3/2 and Co 2p 1/2 due to the spin-orbit coupling. Furthermore, the peak positions of Co 2p 3/2 and Co 2p 1/2 in N-CoO were higher than those in CoO, suggesting that the N incorporation could effectively tune the electronic structure of Co.  Figure 4a displays the LSV curves of the CoO and N-CoO nanowire arrays, with a state-of-the-art Pt/C catalyst as the control sample. Clearly, the N-CoO nanowire arrays displayed a lower onset potential and higher current density than the CoO nanowire arrays, suggesting that the N-CoO nanowire arrays could effectively promote the catalytic activity for HER in the alkaline solution. The overpotential of the N-CoO nanowire arrays at 10 mA cm −2 needed only 123 mV, which was apparently much lower than that of CoO (285 mV, Figure 4c). Moreover, the Tafe slope is an important parameter that can reflect decreased from 26.4% to 22.3%, suggesting the incorporation of N atoms in the lattice of CoO and the formation of oxygen vacancies. The peak of O2 at around 530.1 eV could belong to the OH − . The variation tendency of O2 was also reduced, indicating that CoO improved the desorption of OH − after the incorporation of N, which was beneficial for water dissociation in the alkaline solution. The peak of O3 was considered to be adsorbed water molecules on the surfaces of the CoO and N-CoO nanowire arrays. Inversely, the percentage of O3 in N-CoO significantly increased, implying the enhanced capacity of adsorbing water molecules on N-CoO, which was beneficial for the first step of HER in the alkaline solution. The peak of O4 was regarded as oxygen vacancy, which was caused by the reducibility of NH 3 at high temperature. In a word, the XPS analysis indicated that the electronic structure of the CoO nanowire arrays was tuned significantly after introducing N in CoO, which benefitted the HER in the alkaline solution. Figure 4a displays the LSV curves of the CoO and N-CoO nanowire arrays, with a state-of-the-art Pt/C catalyst as the control sample. Clearly, the N-CoO nanowire arrays displayed a lower onset potential and higher current density than the CoO nanowire arrays, suggesting that the N-CoO nanowire arrays could effectively promote the catalytic activity for HER in the alkaline solution. The overpotential of the N-CoO nanowire arrays at 10 mA cm −2 needed only 123 mV, which was apparently much lower than that of CoO (285 mV, Figure 4c). Moreover, the Tafe slope is an important parameter that can reflect the effects of N-doping on HER kinetics. As shown in Figure 4b, the Tafe slope of the CoO nanowire arrays was 160 mV dec −1 , suggesting that the Volmer step of water dissociation was the rate-determining step on the catalyst of the CoO nanowire arrays. However, the N-CoO nanowire arrays presented a significantly decreased Tafe slope of 97 mV dec −1 , implying the sluggish water dissociation step was accelerated after the incorporation of N in the CoO nanowire arrays. Although the overpotential and Tafe slope of the N-CoO nanowire arrays (123 mV, 97 mV dec −1 ) were still far from the state-of-the-art Pt/C catalyst (49 mv, 41 mV dec −1 ), this design strategy provides a simple method to promote water dissociation in alkaline solutions at the atomic level. Further performance improvements of N-CoO nanowire arrays could perhaps include other transition metals in order to better promote water dissociation.

Results
To evaluate the stability of the N-CoO nanowire arrays, long-term stability measurements of chronoamperometric and polarization cycling were performed. As seen in Figure 4d, both of them displayed no obvious decrease, indicating the excellent electrochemical stability of the N-CoO nanowire arrays.
To assess the energy barriers, we studied the effect of temperature on the performance of the CoO and N-CoO nanowire arrays and found the rate constants followed the Arrhenius relationship. The Arrhenius plots for the CoO and N-CoO nanowire arrays allowed us to extract electrochemical activation energies for HER (Figure 5a) [51]. It shows that the N-CoO nanowire arrays exhibited an apparent barrier value of 18.0 kJ mol −1 , which was significantly lower than that of CoO (25.9 kJ mol −1 ). The significantly decreased electrochemical activation energy of N-CoO for HER was consistent with our XPS analysis, which showed that the N-CoO nanowire arrays could boost water dissociation.
To uncover the electro-transfer reaction kinetics, electrochemical impedance spectroscopy (EIS) was measured. Figure 5b displays the Nyquist plots of the CoO and N-CoO nanowire arrays; both of them displayed semi-circles, which could be well-fitted using a simple equivalent circuit composed of electrolyte resistance (Rs), a constant phase element (CPE) and charge-transfer resistance (R CT ). The charge-transfer resistance (Rct) of the N-CoO nanowire arrays was calculated to be 38.4 Ω, which was much lower than that of CoO (239.7 Ω), further revealing the much-promoted charge transport kinetics of N-CoO nanowire arrays [52].
Considering the importance of surface area on the HER performance, the electrochemical surface areas of CoO and N-CoO nanowire arrays were obtained from the cyclic voltammograms at the scan rates of 20, 40, 60 and 80 mv s −1 . Based on the CVs of Figure S3, the ECSA value of N-CoO was 39.4 mF cm −2 (Figure 5c), which was obviously larger than of N-CoO nanowire arrays could perhaps include other transition metals in order to better promote water dissociation.
To evaluate the stability of the N-CoO nanowire arrays, long-term stability measurements of chronoamperometric and polarization cycling were performed. As seen in Figure  4f, both of them displayed no obvious decrease, indicating the excellent electrochemical stability of the N-CoO nanowire arrays. To assess the energy barriers, we studied the effect of temperature on the performance of the CoO and N-CoO nanowire arrays and found the rate constants followed the Arrhenius relationship. The Arrhenius plots for the CoO and N-CoO nanowire arrays allowed us to extract electrochemical activation energies for HER (Figure 5a) [51]. It shows that the N-CoO nanowire arrays exhibited an apparent barrier value of 18.0 kJ mol −1 , which was significantly lower than that of CoO (25.9 kJ mol −1 ). The significantly decreased electrochemical activation energy of N-CoO for HER was consistent with our XPS analysis, which showed that the N-CoO nanowire arrays could boost water dissociation.
To uncover the electro-transfer reaction kinetics, electrochemical impedance spectroscopy (EIS) was measured. Figure 5b displays the Nyquist plots of the CoO and N-CoO nanowire arrays; both of them displayed semi-circles, which could be well-fitted using a To understand the origin of the high HER performance, the turnover frequency (TOF) of the CoO and N-CoO nanowire arrays was estimated to reveal the intrinsic catalytic activities. The details concerning the obtainment of the TOF values can be found in the supporting information. As shown in Figure 5d, although the surface area of the N-CoO nanowires was obviously larger than that of CoO, the TOF values of the N-CoO nanowires were still significantly larger than that of the CoO nanowires at the same potential, clearly revealing that the incorporation of N in CoO can effectively promote the intrinsic HER activity of N-CoO nanowires [53].
(TOF) of the CoO and N-CoO nanowire arrays was estimated to reveal the intrinsic catalytic activities. The details concerning the obtainment of the TOF values can be found in the supporting information. As shown in Figure 5d, although the surface area of the N-CoO nanowires was obviously larger than that of CoO, the TOF values of the N-CoO nanowires were still significantly larger than that of the CoO nanowires at the same potential, clearly revealing that the incorporation of N in CoO can effectively promote the intrinsic HER activity of N-CoO nanowires [53].

Experiment
All of the chemical reagents were analytical grade (AR) and they were all used directly without any further purification. The deionized water (DI) used in all experiments was produced from the water purification apparatus of Milli-Q (18.2 MΩ cm −1 ).

Experiment
All of the chemical reagents were analytical grade (AR) and they were all used directly without any further purification. The deionized water (DI) used in all experiments was produced from the water purification apparatus of Milli-Q (18.2 MΩ cm −1 ). (Co 2 (NO 3 ) 2 ·6H 2 O), Urea (H 2 C 2 O 4 ), (NH 4 F), (C 3 H 6 O), and (CH 3 CH 2 OH) were purchased from Aladdin Reagent Co., LTD. Carbon was purchased from the Taiwan Carbon Energy Co. LTD.

The Preparation of the CoO and N-CoO Nanowire Arrays
Initially, 40 mmol Co 2 (NO 3 ) 2 ·6H 2 O, 200 mmol H 2 C 2 O 4 and 80 mmol NH 4 F were dissolved in 200 mL deionized water with the help of a magnetic stirring apparatus. Next, a 20 mL solution of the aforementioned mixture and a 2 × 3 cm 2 CC pretreated with acetone, ethyl alcohol and deionized water were put in a Teflon-lined stainless autoclave. The autoclave was transferred to an oven at 200 • C for 6 h. After the temperature of the autoclave reduced to room temperature, the CC was taken out, cleaned using the deionized water and dried in a vacuum oven at 50 • C for 12 h. Then, the precursor of the Co x (OH) y nanowire arrays was prepared by annealing the CC in a muffle furnace at 350 • C for 2 h. Finally, the CoO nanowires were obtained by annealing the Cox(OH)y nanowire arrays at 350 • C in a tube furnace with an argon flow rate of 80 mL/min. The N-CoO nanowire arrays were obtained by annealing the Co x (OH) y nanowire arrays at 320 • C in a tube furnace under a NH3 atmosphere with a flow rate of 60 mL/min.

Materials Characterization
A scanning electron microscope (SEM) and transmission electron microscopy (TEM) were used to obtain the morphology of the Co x (OH) y , CoO and N-CoO nanowire arrays. XRD was applied to provide the crystal structure of the Co x (OH) y , CoO and N-CoO nanowire arrays with a Cu Kα as a source (1.54056 Å). An XPS analysis was performed to distinguish the electronic structure of the CoO and N-CoO nanowire arrays.

Electrochemical Measurements
All the electrochemical tests were performed in 1.0 M KOH with a classic threeelectrode system on an IvumStat electrochemical workstation. The CoO and N-CoO nanowire arrays on CC were employed as working electrodes, and the graphite rod and HgO were used as the counter and the reference electrodes, respectively. All the potentials in the work were converted to the values relative to the reversible hydrogen electrode. The linear sweep voltammetry (LSV) curves shown in the work were collected with a scan rate of 2 mv s −1 . The stability test for N-CoO nanowire arrays was carried out at the potential of −0.123 V. The overpotential of 50 mV with a perturbation amplitude of 10 mV was performed in the measurements of electrochemical impedance spectroscopy (EIS) in a frequency from 10 kHz to 0.1 Hz. The cyclic voltammograms (CVs) in a range between 0 and 0.1 V were used to estimate the electrochemical surface areas (ECSA) at the scan rates of 20, 40, 60, and 80 mV s −1 , respectively. The difference values of current density per square centimeter (cm 2 ) at 0.05 V vs. RHE were plotted with the scan rates in which the slope was used to measure the ECSA. The TOF values were estimated according to the previous literature [54].
For temperature-dependent measurements, the electrochemical reaction cell with electrodes was put in a water bath. The temperature was first improved to a value above 60 • C, then the current values of the catalyst at temperatures of 55, 45, 35 and 25 • C were recorded with the temperature reducing in the environment in order to decrease error. As we all know, the kinetics of chemical reactions increase with rises in temperature which is the same for HER. The relationship between the chemical rate constant and temperature is roughly proportional to exp(−∆G/kT), where ∆G and k are the activation energy and the Boltzmann constant, respectively. Specifically, the approximate activation energy (Ea) for HER can be defined by the Arrhenius relationship where T is the test temperature, ik is the current density at an overpotential of −200 mV and R is the gas equilibrium constant. The Ea for different HER catalysts can be obtained by the slopes of various Arrhenius plots (-R×slope).

Conclusions
In general, we have demonstrated that the incorporation of N in CoO nanowires can effectively improve the HER performance in an alkaline solution. Compared with the inert HER activity of CoO nanowires, the obtained N-CoO nanowires obviously exhibited increased HER activity, which needed only 123 mV to achieve the current density of 10 mA cm −2 . The XPS of Co 2p and O 1s revealed that the electronic structure of CoO was obviously modulated after the incorporation of N, which improved the adsorption and activity of the water molecule. The energy barriers obtained from the Arrhenius relationship of the current density at different temperatures indicated that the N-CoO nanowires accelerated water dissociation in the alkaline solution. This work provides a new design insight for the development of Pt-free catalysts with a high performance for HER in alkaline solutions at the atomic scale.